www.biogeosciences.net/bg/1/159/ SRef-ID: 1726-4189/bg/2004-1-159 European Geosciences Union
Biogeosciences
Large-scale environmental controls on microbial biofilms in
high-alpine streams
T. J. Battin1,2, A. Wille3, R. Psenner3, and A. Richter4
1Department of Limnology, IECB, University of Vienna, A-1090 Vienna, Austria 2Department of Ecology, University of Barcelona, 08028 Barcelona, Spain
3Institute of Zoology and Limnology, University of Innsbruck, A-6020 Innsbruck, Austria 4Department of Chemical Plant Physiology, IECB, University of Vienna, A-1090 Vienna, Austria Received: 26 July 2004 – Published in Biogeosciences Discussions: 31 August 2004
Revised: 6 December 2004 – Accepted: 16 December 2004 – Published: 27 December 2004
Abstract. Glaciers are highly responsive to global warming and important agents of landscape heterogeneity. While it is well established that glacial ablation and snowmelt regu-late stream discharge, linkage among streams and streamwa-ter geochemistry, the controls of these factors on stream mi-crobial biofilms remain insufficiently understood. We in-vestigated glacial (metakryal, hypokryal), groundwater-fed (krenal) and snow-fed (rhithral) streams – all of them repre-sentative for alpine stream networks – and present evidence that these hydrologic and hydrogeochemical factors differ-entially affect sediment microbial biofilms. Average micro-bial biomass and bacterial carbon production were low in the glacial streams, whereas bacterial cell size, biomass, and car-bon production were higher in the tributaries, most notably in the krenal stream. Whole-cell in situ fluorescence hybridiza-tion revealed reduced detechybridiza-tion rates of theEubacteria and higher abundance ofα-Proteobacteriain the glacial stream, a pattern that most probably reflects the trophic status of this ecosystem. Our data suggest low flow during the onset of snowmelt and autumn as a short period (hot moment) of fa-vorable environmental conditions with pulsed inputs of al-lochthonous nitrate and dissolved organic carbon, and with disproportionately high microbial growth. Tributaries are rel-atively more constant and favorable environments than kryal streams, and serve as possible sources of microbes and or-ganic matter to the main glacial channel during periods (e.g., snowmelt) of elevated hydrologic linkage among streams. Ice and snow dynamics – and their impact on the amount and composition of dissolved organic matter – have a crucial impact on stream biofilms, and we thus need to consider mi-crobes and critical hydrological episodes in future models of alpine stream communities.
Correspondence to:T. J. Battin (tomba@pflaphy.pph.univie.ac.at)
1 Introduction
Alpine glaciers are prominent agents of environmental het-erogeneity. At large temporal scales, glacial retreat shapes landscapes by generating chronosequences of revegetation and humus build-up which in turn alters nitrogen fixation and carbonate leaching – factors that can affect solute fluxes at the catchment level and even lake phytoplankton commu-nities (Engstrom et al., 2000). At smaller temporal scales, the seasonal dynamics of glaciers can trigger different hy-drologic reservoirs, change flowpaths of water and solutes through the catchment, and drastically shape the fluvial geo-morphology of lotic networks (e.g. Malard et al., 1999; Smith et al., 2001). For instance, during summer, supraglacial and subglacial ice-melt feeds glacial streams and replenishes groundwater stores; groundwater becomes increasingly im-portant in autumn and together with subglacial waters con-stitutes the main source to winter flow (e.g. Smith et al., 2001). Each of these hydrological sources generates distinct chemical signatures, which, depending on the level of hydro-logic linkage among streams and terrestrial/aquatic connec-tivity, can induce remarkable physical and geochemical het-erogeneity at the landscape level (e.g. Tockner et al., 1997; Smith et al., 2001; Brown et al., 2003). Along with this shift-ing dominance of hydrologic flowpaths, seasonal variation in discharge also leads to cycles of expansion and contrac-tion of the lotic network (Malard et al., 1999). Meltwaters from glacial ablation and snowmelt can dramatically increase discharge, expand channel networks and increase hydrologic connectivity whereas receding discharge in autumn and win-ter contracts networks.
sedi-Fig. 1. Map of the Rotmoos catchment with the location of the study streams. Upstream sites include the hypokryal and rhithral, downstream sites include the metakryal and krenal. Contour lines refer to the altitude (meters above sea level).
ment loads, low channel stability and large diel fluctuations. So-called rhithral streams are largely fed by snowmelt and characterized by clear seasonal dynamics. By contrast, kre-nal streams fed by groundwater are generally more stable. To understand how this heterogeneity affects the environmental controls (i.e. physical and chemical), and how, in turn, they interact to determine the habitat template for benthic com-munities, it is essential to consider the range and scales of links between environmental factors and the stream ecosys-tem (Brown et al., 2003). These authors proposed a con-ceptual model that links environmental variables from the regional (e.g. climate), to the catchment (e.g. glacier and snow dynamics), stream reach (e.g. hydrochemistry, flow) and patch scale. Similarly, Milner et al. (2001) related lon-gitudinal gradients in macroinvertebrate community struc-ture to channel stability and maximum water temperastruc-ture – both being a function of the distance from the glacier and time since deglaciation. Habitat heterogeneity was also pro-posed to mitigate glacial-induced disturbances, and to in-crease macroinvertebrate biodiversity and community stabil-ity at the floodplain level (Burgherr et al., 2002). Streams that are less prone to the glacial meltwaters and snowmelt pulses may serve as refugia for macroinvertebrates and post-disturbance sources of recolonization for the main glacial channel (Burgherr et al., 2002).
Seasonal variation is also pronounced in glacial streams
with snow cover and glacial flow pulses as major agents of environmental harshness, and low flow conditions in early spring and autumn can act as relatively short periods of more favorable conditions. Such windows of opportunity (Mil-ner et al., 2001) or hot moments (McClain et al., 2003) are characterized by disproportionately high reaction and activ-ity rates, and are thought to greatly contribute to the function-ing of glacial stream ecosystems. Despite the awareness that alpine aquatic ecosystems are particularly prone to climate change (Chapin and K¨orner, 1994; McGregor et al., 1995), the response of this predictable seasonal variation to global warming and its consequences for community and ecosystem ecology remain poorly understood.
While considerable effort has addressed the influence of glacial dynamics on algal and especially macroinvertebrate communities (Hieber et al., 2001; Peterson et al., 2001; F¨ureder et al., 2001; Burgherr et al., 2002), virtually nothing is known on the microbial ecology of alpine streams (Bat-tin et al., 2001; Logue et al., 2004). The central objective of the present paper is to study the variability of stream mi-crobial biofilms within a glacial floodplain and to assess the influence of glacial ablation and snowmelt as major hydro-logic factors determining the physical template for stream biofilms. We selected kryal, rhithral and krenal streams in a small glacial catchment in the Austrian Alps to study dif-ferent impacts of glacier and snow dynamics on the hydrol-ogy and biogeochemistry of these streams depending on their distance from the glacier and their position in the catchment. We hypothesized that this combination of linkage and loca-tion dictates spatial and temporal varialoca-tion of bulk microbial biomass and activity, cell size structure and microbial com-munity composition.
2 Study sites
Study sites were located within the glacial catchment (Rot-moos, 11◦03′N 46◦50′E, 2280 to 2450 m above sea level) in the Austrian Alps (Fig. 1). Approximately 40% of the catchment area (10 km2)are glaciated. Annual mean tem-perature is−1.3◦C (1997/1998) and annual precipitation
Table 1. Characterization of the Rotmoos study sites and bulk features of the sediment collected for microbial analyses. Given are means±SE.
Site Location Stream typea Altitude Median grain size Bulk density Organic matter (m a.s.l.)b (µm) (g cm−3) (%w/w) Site 1 Main channel Metakryalc 2360 514±84 1.665±0.074 0.02±0.03 Site 2 Tributary Rhithral 2320 467±38 1.678±0.034 0.18±0.05 Site 3 Main channel Hypokryald 2250 355±30 1.505±0.027 0.06±0.02 Site 4 Tributary Krenal 2255 436±65 1.599±0.056 0.19±0.01
aaccording to Ward (1994)
bmeters above sea level c0.3 km from glacial terminus
d3.0 km from glacial terminus
with grains of carbonate) and fluvio-glacial sands. The entire study area lies above tree line. Vegetation is characterized by pioneering plants such asSaxifraga aizoides, Saxifraga op-positifoliaor Cerastium uniflorum, and acidic alpine grass-lands containingCarex sempervirens(Kaufmann, 2001).
We selected one upstream (metakryal) and one down-stream (hypokryal) study site within the main channel (Fig. 1, Table 1). The rhithral site is largely fed by meltwater from the metacarbonate subcatchment and located between the 1920 and 1994 terminal moraine. The groundwater-fed krenal stream is located within the peat deposit, and water ex-periences sensitive downstream warming due to solar radia-tion. Bulk sediment density ranged from 1.50 to 1.68 g cm−3, and sediment organic matter was higher in the tributaries than in the glacial stream (Table 1). Major algae associated with sediments wereHydrurus foetidusin the glacial stream and representatives of the diatom generaFragilaria, Syne-dra,Cymbella,Diatoma,NaviculaandNitzschiadominated the tributary samples.
3 Sample collection
We collected random grab-samples of benthic sediments (at a depth of 0 to 4 cm) on 9 to 12 occasions from September 1998 to December 1999 at all four sites. Avalanches made access to the sample sites impossible during winter. Sed-iment samples were gently passed through a 1000µm sieve in the field and the fraction>200µm was retained to achieve sandy samples with median grain size from 355 to 528µm. Samples for bacterial secondary production were incubated in the field with radiochemicals to ensure in situ tempera-ture, which varied among streams and with time. Aliquot samples were immediately transferred into sterile polypropy-lene tubes containing 2.5% (v/v) formaldehyde and a 1:1 mixture of ethanol and paraformaldehyde (4%, w/v) for the determination of bacterial biomass and whole-cell in situ hybridization, respectively. The remaining sediment was frozen (−20◦C) for carbohydrate and chlorophyll a
analy-ses, which were performed within 2 months. One-liter sam-ples of stream water were collected in acid rinsed bottles for chemical analyses.
4 Methods
4.1 Flow gauging
Highly unstable channels and logistic constraints prevented continuous flow monitoring in all four study sites. However, discrete flow measurements from the hypokryal site could be related (r2=0.987, slope=1.02, n=27, p<0.001) to flow records from a downstream gauging station (TIWAG, Ober-gurgl). Based on this regression model, we reconstructed the annual hydrograph for the hypokryal site, which we then used to define hydrological periods. An unusually large storm event (49.2 m3s−1at the TIWAG gauging station) that caused massive channel alterations in September 1999 was excluded from the hydrograph reconstruction.
4.2 Stream water chemistry
Conductance was measured in the field with a conductivity meter (WTW LF196, Germany). Anions (SO−4, NO−3, Cl−) were analyzed with ion chromatography (DIONEX DX-120) and cations (Ca2+, Mg2+, Na+, K+)with an atomic absorp-tion spectrometer (Perkin Elmer 3110). Filtered (Whatman, GF/F) samples were analyzed with a Shimadzu TOC-5000 for DOC.
to McKnight et al. (2001) as the ratio of the fluorescence in-tensities measured at 450 nm and 500 nm using an excitation of 370 nm. We used simple two end-member mixing analysis to assess the relative source contribution to the DOC pool at landscape level (cf. Hood et al., 2003). FI values from glacial ice and streamlets draining the peat deposit were used as po-tential end-members.
We computed the Renkonen index (Krebs, 1989) based on ion concentrations (Ca2+, Mg2+, Na+, K+, Cl−, SO−4) to describe the geochemical similarity between pairs of streams (hypokryal and metakryal, rhithral and krenal). We used the relationship of this similarity index with the main chan-nel discharge to evaluate the geochemical linkage among streams and its seasonal variation.
4.3 Bacterial biomass and chlorophylla
Bacterial abundance was estimated by epifluorescence mi-croscopy after staining with 4′, 6′-diamidino-2-phenylindole (DAPI, Sigma, St. Louis USA) according to Porter and Feig (1980). Sediment preserved in formaldehyde (2.5%) was treated with 0.1 mol L−1tetrasodium pyrophosphate and sonicated (180 s, 40 W output) to detach bacterial cells (Velji and Albright, 1985). To reduce the high background fluo-rescence caused by minerals, 2 ml of the thoroughly mixed supernatant were transferred into Eppendorf tubes, sonicated (microtip, 30 s, 30 W output) and centrifuged (Eppendorf 5415C) to pellet mineral particles. This procedure signif-icantly reduced background fluorescence and recovered on average>93% of the bacterial cells, as revealed by micro-scopic analyses of residual particles. Bacterial cells were enumerated on a black 0.2µm GTBP Millipore filter in 10– 30 randomly selected fields to account for 300 to 500 cells. Two filters were counted per site and date. Cell size and shape were determined on 400 to 600 bacterial cells per filter using image analysis as described by Posch et al. (1997). Cell volume was derived fromV=(w2×p/4)×(l−w)+(p×w3/6), whereV is the cell volume (µm3),wandlare cell width and length (µm), respectively. We used the allometric relation-shipC=120×V0.72 between cell volume (V,µm3)and cell carbon content (C, fg) to calculate bacterial biomass (Nor-land, 1993).
Sediment chlorophyll a was extracted with methanol (99%, v/v) from approximately 2–3 g wet sediment during 12 h in the dark (4◦C). After centrifugation, the supernatant was assayed fluorometrically (EX435/EM675) and chloro-phylla from spinach (Sigma, St. Louis, USA) was used as standard.
4.4 Bacterial secondary production
The incorporation of [3H]leucine was used to estimate sed-iment bacterial production (Marxsen, 1996). Approxi-mately 2–3 g (wet weight) sediment were incubated with [3H]leucine (specific activity 60 Ci mmol−1, American
Ra-diochemicals) at saturating concentrations of 200 nmol L−1 (made up with cold leucine) for 2 to 4 h in the field at in situ temperatures in the dark. Time series confirmed linear [3H]leucine uptake during this incubation time. Triplicate [3H]leucine assays and duplicate formaldehyde-killed con-trols (2.5% final concentration, 30 min prior to [3H]leucine addition) were employed per run. The [3H]leucine incorpo-ration was terminated in the field with 5 ml formaldehyde and samples were frozen (−20◦C) within 2 to 3 h. Upon return
to the laboratory, samples were repeatedly washed with 5% formaldehyde, and proteins were extracted, precipitated and radioassayed as described elsewhere (Battin et al., 2001). 4.5 Whole-cell in situ hybridization
Whole-cell fluorescence in situ hybridization (FISH) was performed to describe the community composition of biofilm samples from 12 August, 6 September, and 3 October 1999. Within 3 days after sampling, bacterial cells were detached from the sediment as described above. Aliquot samples (3 to 5 ml) of the supernatant were filtered onto 0.2µm poly-carbonate membrane filters (Millipore GTTP, 47 mm diam-eter) and subsequently washed with phosphate buffer saline (PBS, pH 7.2). The following oligonucleotide probes (Inter-activa, Ulm, Germany) were used as in Battin et al. (2001): ARCH915 for members of the domainArchaea(16S rRNA, positions 915 to 934), EUB338 for members of the do-mainBacteria (16S rRNA, positions 338 to 355), BET42a for the beta subclass of Proteobacteria (23S rRNA, posi-tions 1027 to 1043), ALF968 for the alpha subclass of Pro-teobacteria(16S rRNA, positions 968 to 986) and CF319a for theCytophaga-Flavobacterium cluster (16S rRNA, po-sitions 319 to 336). Probes were labeled with the indocar-bocyanine fluorescent dye CY3 (Biological detection Sys-tems, Pittsburgh, PA USA). The unlabeled probe GAM42a served as competitor for BET42a. Hybridization and epiflu-orescence microscopical (Axioplan Carl Zeiss Inc.) counts of hybridized and DAPI stained cells were performed as pre-viously described (Alfreider et al., 1996; Battin et al., 2001). As the fluorochrome CY3 sorbed to minerogenic particles in some samples, thereby increasing the background fluo-rescence, each particle emitting a CY3 signal was checked against DAPI staining. We also checked each filter section for autofluorescence signals of phototrophic and cyanobac-terial cells using the filter sets 510–560 (FT580, LP590) and 450–490 (FT510, LP520).
4.6 Microbial exopolysaccharides
Fig. 2. Stream annual hydrograph (hypokryal stream) illustrating hydrological periods characteristic of glacial streams. The line de-notes the hydrograph, dots refer to sampling campaigns.
at−20◦C for 48 h. The polymeric fraction was pelleted by
centrifugation (2500 rpm, 30 min) and redissolved in MilliQ water. EPS was analyzed for bulk exopolysaccharides with the sulfuric acid – phenol method (Dubois et al., 1956).
4.7 Statistical analyses
Repeated-measures analysis of variance (rmANOVA) was used to test the influence of the hydrological period (win-ter flow, snow melt, ablation, recession) as the repeated vari-able and stream type (metakryal, hypokryal, rhithral, krenal) on an array of dependent variables related to streamwater chemistry and biofilm structure and function. Hydrologi-cal periods (see Fig. 2) were identified using the multivariate Wilk’sλinstead of the univariate statistics where data did not meet sphericity (Max and Onghena, 1999). Tukey’s multiple comparison post-hoc test followed rmANOVA to asses how levels of stream type differ. Variation in selected microbial parameters was explored with stepwise forward multiple re-gression on data pooled from all stream types to account for catchment-scale variation and to achieve degrees of freedom high enough to ensure robust results. Entrance criterion was P=0.15 and minimum tolerance for entry into the model was 0.01. We computed principal component analyses (PCA) with the median of cell length, width, volume and circularity to explore bacterial cell size structure. Community composi-tion was also analyzed using PCA on the relative abundance of the various FISH probes. We used correlation matrices for the PCA and 2 factors were retained for analysis. Data were natural-log or arcsin-square-root transformed prior to analy-sis. All data are given as mean±standard error (S.E.) and
sta-tistical tests were considered significant at the levelα=0.05. All analyses were performed with SYSTAT (Evanston, SYS-TAT, Inc.).
Table 2. Summary statistics and repeated measures ANOVA for streamwater chemistry. Site (metkryal, rhithral, hypokryal, krenal) is the independent variable, hydrological periods (winter flow, snow melt, ablation, recession) are the repeated variables.
Sites ANOVA F-statisticsa
Parameter Metakryal Rhithral Hypokryal Krenal Stream Hydrologic period Interaction
Temperature (◦C) 1.32±0.29 3.91±1.11 3.60±0.70 4.42±1.19 9.386* 32.627*** 2.163ns Conductance (µS cm−1) 115±23 186±14 114±13 70±4 46.784** 27.321*** 7.671** Ca2+(mg L−1) 13.87±2.55 23.84±1.82 14.16±1.64 6.70±0.54 75.483*** 41.065*** 7.356** Mg2+(mg L−1) 3.21±0.91 5.68±0.71 2.75±0.45 1.54±0.17 42.776** 24.273*** 4.453** K+(mg L−1) 1.61±0.36 2.53±0.23 1.69±0.20 1.06±0.06 32.769** 60.772*** 16.212*** Na+(mg L−1) 0.47±0.12 0.52±0.11 0.43±0.09 0.61±0.09 0.439ns 5.998** 1.056ns SO−4(mg L−1) 19.09±3.93 18.52±2.22 14.90±1.75 12.52±0.71 1.978ns 32.890*** 11.313*** Cl−(mg L−1) 0.30±0.07 0.45±0.08 0.31±0.07 0.33±0.12 12.355** 27.659*** 3.048* NO−3(mg L−1) 0.67±0.16 0.29±0.09 0.60±0.11 0.56±0.12 30.997** 188.72*** 9.921** DOC (mg L−1) 0.32±0.03 1.61±1.22 0.61±0.13 0.54±0.07 0.673ns 0.939ns 0.324ns FIc 1.71±0.08 1.58±0.03 1.56±0.04 1.52±0.04 NAb NA NA
ans denotesP >0.05, *P <0.05, **P <0.01, ***P <0.001 bNA, not analyzed because of missing values
cdimensionless index
Fig. 4. Hysteresis curves generated from pairwise comparisons (Renkonen Similarity index) of the hypokryalversusmetakryal (a) and rhithralversuskrenal (b) and based on streamwater ion concen-trations. REC: recession, WIN: winter flow, SNO: snowmelt, ABL: ablation.
5 Results
5.1 Hydrogeochemistry
Four distinct periods characterized the reconstructed hydro-graph in the hypokryal stream (Fig. 2). Flow was low and constant during winter snow cover, dramatically increased during snowmelt in spring, high and variable during ablation in summer, and receded in autumn.
Most geochemical descriptors differed significantly
among streams and hydrologic periods, and with signifi-cant interaction terms (i.e., time x site interaction,P <0.05, Table 2). Streamwater conductance (rmANOVA, Wilk’s λ, P=0.03) and calcium concentration (rmANOVA, Wilk’s λ, P=0.003) – as general geochemical descriptors – were significantly elevated at the onset of snowmelt, decreased af-ter snowmelt, were very low during ablation and increased again during the recession period (Table 2, Fig. 3). The metacarbonate outcrop in the southern subcatchment im-parted distinct signatures with higher concentration of cal-cium (ANOVA, Tukey P <0.001), magnesium (ANOVA, TukeyP <0.01) and potassium (ANOVA, TukeyP <0.001) to the rhithral stream. This effect was attenuated in both the metakryal and the hypokryal and was virtually absent from the krenal site (Table 2).
NO3-N concentration showed significant (rmANOVA, Wilk’s λ, P=0.02) temporal patterns in all sites with high concentrations at the beginning of the snowmelt that im-mediately decreased to reach minimal values during glacial ablation (Table 2, Fig. 3). Average DOC concentrations exhibited no significant seasonal trend (rmANOVA, Wilk’s λ, P >0.05), except a concentration peak in the hypokryal site at the onset of the snowmelt (Table 2, Fig. 3). Av-erage DOC concentrations were lower in the sites close to the glacier forefield (metakryal: 0.32 mg C L−1, rhithral: 0.39 mg C L−1) than in the downstream sites adjacent to the peat deposit (hypokryal: 0.60 mg C L−1, krenal: 0.55 mg C L−1).
Table 3.Correlation between discharge and selected chemical pa-rameters in the hypokryal site (n=14, except for FI: n=10).
Dependent variable Pearson’sr P
Conductance −0.839 <0.001 Ca2++ Mg2+ −0.841 <0.001
Cl− −0.725 0.003
Na+ −0.839 0.023
K −0.601 <0.001
NO3-N −0.512 0.074 DOC −0.497 0.100 FI −0.696 0.012 % glacial DOC −0.692 0.025
higher than from both the hypokryal and krenal. FI values peaked during snowmelt and the beginning of the ablation in the hypokryal and metakryal, respectively. FI values of shallow glacial ice averaged 1.92±0.16 with maximum
val-ues as high as 2.20 during a bloom ofChlamydomonas ni-valis,whereas the FI values of peat draining waters averaged 1.29±0.13. Mixing model analysis using these FI values as
end-members suggests that 62±7% of the fulvic acids in the
metakryal originate from glacial ice, a relative contribution that decreased to 43±8% in the hypokryal. The krenal stream had lowest contributions (33±6%) of glacier-derived fulvic acids.
Stream discharge correlated significantly with most chem-ical parameters in the hypokryal with high ionic concentra-tions during low flow (Table 3). While we found no such correlation with DOC concentration, FI and the percentage contribution of glacial-derived fulvic acids were inversely re-lated to discharge.
Clear hysteresis patterns result from plots of the Renkonen similarity index from pairwise comparison of streamwater geochemistry and discharge (Fig. 4). The geochemical fin-gerprint became increasingly similar between the hypokryal and metakryal as winter flow progressed, decreased with the onset of snowmelt and reached minimal similarity during re-cession. Pairwise comparisons between the rhithral and kre-nal give similar results.
5.2 Spatial and temporal patterns of microbial biomass and activity
Average microbial biomass and activity ranged 37 and 14-fold, respectively, among stream types and 3 and 2-14-fold, re-spectively, among hydrologic periods, and the influence of stream type remained consistent across hydrologic periods (i.e., time x site interaction not significant, P >0.05) (Ta-ble 4, Fig. 5). Chlorophyllawas highest in the krenal stream (ANOVA, Tukey P <0.05), and significantly (rmANOVA, Wilk’sλ, P=0.016) changed with hydrological periods with low values during both winter flow and glacial ablation, and
Fig. 5. Temporal variation of various sediment biomass parame-ters and bacterial carbon production (BCP) in the glacial stream sites and the tributaires. The arrow indicates the onset of the snowmelt and the gray shading indicates glacial ablation. Data are means±SE.
Table 4.Summary statistics (mean±SE) and repeated measures ANOVA for microbial biofilm parameters. Site (rhithral, krenal, metakryal, hypokryal) is the independent variable, hydrological periods (winter flow, snow melt, ablation, recession) is the repeated variable.
Site ANOVAF-statisticsaParameter Metakryal Rhithral Hypokryal Krenal Site Hydrologic period Interaction
Chlorophylla(µg g−1DM) 0.01±0.01 0.45±0.16 0.02±0.01 0.86±0.39 41.05*** 11.30*** 0.496ns
Bacterial abundance (106cells g−1DM) 4.38±0.99 96.3±21.4 18.70±4.76 154±34 111.88** 8.94** 0.489ns
Cell length (µm) 0.94±0.07 0.85±0.03 0.81±0.03 0.89±0.02 0.974ns 5.915* 1.146ns
Cell volume (µm3) 0.11±0.02 0.08±0.01 0.08±0.01 0.10±0.01 1.419ns 10.485** 0.980ns
Cell circularityb 0.80±0.001 0.83±0.001 0.82±0.001 0.82±0.001 3.015ns 3.459* 1.603ns
Cell C-content (fg C cell−1) 18.35±1.32 17.17±0.83 15.58±0.66 21.24±1.60 12.05* 9.98** 2.31ns
Bacterial biomass (µg C g−1DM) 0.080±0.001 1.68±0.39 0.287±0.075 2.95±0.53 103.91*** 8.23** 0.853ns
BCP (µg C h−1g−1DM) 0.03±0.01 0.41±0.07 0.15±0.03 0.34±0.07 62.38*** 3.33ns 1.05ns
Exopolysaccharides (µg C g−1DM) 1.32±0.19 1.83±0.33 2.29±0.40 4.60±0.55 11.45* 1.55ns 1.20ns
Exopolysaccharides: cell (pg C cell−1) 0.66±0.14 0.03±0.01 0.25±0.06 0.4±0.01 32.59** 5.25* 0.441ns
ans denotesP >0.05, *P <0.05, **P <0.01, ***P <0.001 bdimensionless
Table 5. Detection rates (%) of fluorescence in situ hybridization (FISH) relative to total counts of DAPI-stained bacterial cells. Data are given as average±SE, (n=3).
Site Eubacteria α β Cytophaga- Archaea
Proteobacteria Proteobacteria Flavobacterium
Metakryal 31.33±7.43 3.73±0.69 16.67±1.77 1.87±1.67 2.80±1.48 Rhithral 63.33±2.96 1.80±0.47 22.67±4.18 1.97±1.22 2.33±0.61 Hypokryal 41.67±2.61 2.90±0.15 19.67±1.20 1.00±0.44 1.70±1.11 Krenal 60.00±6.25 1.83±0.48 15.67±3.28 2.13±0.58 2.27±0.96
the onset of snowmelt and crashed during ablation in the hy-pokryal. This temporal variation was less clear in the tribu-taries (Fig. 5) and did not differ significantly among hydro-logical periods (rmANOVA, Wilk’sλ, P >0.05).
Bacterial carbon production significantly differed among stream types with highest values in the tributaries (rmANOVA followed by Tukey, P <0.05) (Table 5). Al-though our estimates of bacterial carbon production are few during winter, there is evidence for a peak at the onset of the snowmelt in the hypokryal site (Fig. 5). Yet rmANOVA did not reveal any significant effect of hydrological period on bacterial carbon production (Table 4). Chlorophyll a, bacterial biomass and DOC FI values accounted for 77% (F=27.54, P <0.001) of the temporal and spatial variation in bacterial carbon production. Water temperature did not improve this regression model (data not shown).
Average concentrations of exopolysaccharides were high-est in the krenal tributary (ANOVA, Tukey P <0.001), whereas average exopolysaccharide to cell ratios were highest in the metakryal (ANOVA, Tukey P <0.001) (Ta-ble 4). Exopolysaccharide to cell ratios differed significantly (rmANOVA, Wilk’sλ, P=0.014) among hydrologic periods with lowest values during snowmelt and ablation and el-evated values during low flow in autumn and winter (Ta-ble 4, Fig. 5). Multiple regression analysis revealed that calcium concentration and bacterial carbon production
ex-plained 36% (F=10.59,P <0.001) of the spatial and tempo-ral variation in exopolysaccharide to cell ratios.
Principal component analysis clearly grouped hydrolog-ical periods based on morphologhydrolog-ical parameters of bacte-rial cells (Fig. 6a). Median cell length and volume loaded strongly on PC1 (component loading: 0.93) illustrating the differences between winter flow and snowmelt versus abla-tion and recession. Whereas median cell circularity loading on PC2 (component loading: 0.80) indicated a shift from more rod-shaped cells during recession and winter flow to small and coccoid cells during snowmelt and glacial abla-tion. Representative plots of cell length frequency during winter flow and ablation illustrate this shift in cell morphol-ogy in each of the stream types (Fig. 6b). Repeated-measures ANOVA (Wilk’sλ, P <0.05) confirmed the statistical signif-icance of the temporal course of these morphological cell pa-rameters (Table 4).
5.3 Community composition
Fig. 6.Principal component analysis of the bacterial cell size structure(a)and representative frequency analyses of bacterial cell length for ablation and winter flow(b).
hypokryal sites. These differences were statistically sig-nificant for the Eubacteria (ANOVA, P=0.009) and the α-Proteobacteria (ANOVA, P=0.015). Probes for the β -Proteobacteria,Cytophaga-Flavobacterium cluster and the Archaeadid not reveal any site-specific pattern. The tem-poral shift in community composition was shaped by the September storm that increased the numbers of Cytophaga-Flavobacterium andArchaeaprobes in all streams but did not affect the partitioning effect ofα-Proteobacteria.
6 Discussion
Our results reveal snowmelt and glacial meltwaters as im-portant environmental agents that shape the habitat template in alpine streams and, in turn, affect the structure and func-tion of microbial biofilms. Both effect and strength of this template operating at the reach scale depend on large-scale patterns such as landscape positioning of the streams relative to the glacier, snowpack, and geology.
Landscape positioning and linkage among streams (i.e., induced by hydrology) were clearly reflected by the hydro-geochemistry – as suggested by the significant interaction terms of the repeated measures ANOVA (Table 2). For in-stance, the metacarbonate outcrops increased ion concentra-tions in the rhithral, to some extent in the metakryal and with an even lower signature in the downstream hypokryal. This was particularly pronounced during low winter flow and be-came less apparent during ablation. Conversely the krenal stream had lower ion concentrations throughout the entire
Fig. 7. Principal component analysis of the data of whole cell fluorescencein situ hybridization on samples from 12 August, 6 September and 3 October. Arrows indicate the temporal trajectory of microbial communities. Data are from the metakryal (λ), hy-pokryal (ϒ ), rhithral (ν)and krenal stream (o).
in-stance, discharge from small nival tributaries draining ele-vated sub-catchments are insignificant during this hydrolog-ical period when streamwater largely originates from sub-glacial and groundwater sources (cf. F¨ureder et al., 2001; Smith et al., 2001). This contrasts with snowmelt and abla-tion as periods of higher hydrological linkage that reduce the geochemical similarity between the hypokryal and metakryal sites. Depending on landscape positioning various hydrolog-ical reservoirs are progressively activated during these peri-ods, produce water parcels with different geochemical fin-gerprints that enter the glacial streams along multiple flow-paths. These patterns agree with other glacial catchments, where shifts from groundwater to surface water dominated ecosystems result in considerable chemical heterogeneity at the landscape level (e.g. Malard et al., 2000).
Furthermore, spatial patterns of DOC fluorescence proper-ties strongly suggest accompanying terrestrial vegetation as a significant carbon source to the streams. While the elevated FI values in the metakryal are evidence for algal/microbial derived precursor material in the upper reaches of the glacial stream, lower FI values in the downstream hypokryal and krenal sites indicate increasing inputs of terrestrially-derived fulvic acids. The relationship between discharge and mixing-model results suggest that fulvic acids of terrestrial origin are highest when the terrestrial/aquatic linkage is elevated, and that the glacier constitutes an important DOC source to the stream when the terrestrial/aquatic linkage is reduced during winter. This implies that glaciers and snow field act as col-lectors of airborne organic material, such as leaves, debris, pollen, dust, insects etc. but also as habitats for snow algae and other microorganisms which produce or process organic matter (S¨awstr¨om et al., 2002). Up to now, ice and snow were considered primarily as drivers of hydrologic events, responsible for seasonal discharge dynamics and water tem-perature. Our data now expand this view by showing that DOC (fulvic acids) from these cold habitats adds to the com-plexity of high-alpine streams. Yet care should be taken in in-terpreting yields from end-member mixing analyses because they are based on the assumption that the fluorescence in-dex can be used quantitatively (Hood et al., 2003). Never-theless, the spatial pattern of our estimates agrees with the chronosequence of revegetation and soil build-up from the glacier forefield with almost barren recent glacial sediments to the downstream peat deposit, and with the soil organic car-bon contents along this gradient (cf. Tscherko et al., 2003). These findings also support other reports on soils as DOC sources and longitudinal shifts in DOC chemistry in alpine catchments (e.g., Boyer et al., 1997; Hood et al., 2003).
Our data suggest that snowmelt is a major factor for micro-bial biofilms for several reasons. Snowmelt increases the hy-drologic linkage among streams and their connectivity with the terrestrial milieu via surface runoff and by recharging the groundwater, which can then feed the streams (cf. Smith et al., 2001; Brown et al., 2003). Interactions between ground-water and streams as triggered by snowmelt was recently
shown to be more important in high elevation catchment that previously thought (Liu et al., 2004). These flowpaths reacti-vate and transport resources (DOC, nutrients) and most likely also microbes to the streams. The onset of snowmelt is also characterized by low turbidity and flow-induced disturbance, and is therefore considered as a hot moment (McClain et al., 2003) crucial for rapid biofilm recovery from winter tran-sient state and followed by a period of flow-induced stress caused by increasing snowmelt and glacial ablation. In fact, snowmelt period is characterized by two phases. The initial flush, triggered by the first warming period in spring, delivers small volumes of concentrated solutions of ions, depending on frequent phase transitions (ice, water, vapour) in the snow-pack and the accumulation of airborne material during win-ter. The second phase is characterized by large amounts of meltwater low in solute concentration (Psenner and Nickus, 1986). We found peaks (up to 4-fold) of bacterial abun-dance and biomass in the metakryal and both tributaries shortly after the onset of snowmelt. During snowmelt av-erage bacterial abundance was notably higher in the rhithral (206×106cells g−1DM) and the krenal (234×106cells g−1
DM) than in other alpine streams (e.g. Logue et al., 2004). This suggests – among other factors – inputs of microbial cells from the snowpack. Snow cover was in fact reported to harbor an active and diverse microbial community (e.g. Al-freider et al., 1996; Sattler et al., 2001), yet flushing of soils during snowmelt should be considered as a further source of microbes. Concomitantly with these biomass patterns our data – although few during winter – indicate increased bac-terial carbon production in the hypokryal and krenal streams shortly after the onset of snowmelt, hinting at favorable en-vironmental conditions during this short period. Concur-rently, our data from the hypokryal show peaks of DOC and NO3-N concentrations, probably attributable to snowmelt. Unfortunately, we do not have comparative data from the metakryal because this site is hardly accessible during win-ter. However, pattern congruency of available data supports our view, which is furthermore corroborated by the numer-ous reports from others showing pulsed snowmelt-induced inputs of NO3-N and other solutes (Campbell et al., 1995; Williams et al., 2001; Tockner et al., 2002). Snowpack NO3-N from wet and dry nitrogen deposition can be remarkable (7 to 13 kg N ha−1yr−1)in the Central Alps during the end of snow accumulation when the lower tropospheric boundary layer extends upwards to glacial altitudes (e.g. Psenner and Nickus, 1986).
the source area is expanding as snowmelt occurs at higher al-titudes and thereby affecting upper stream reaches (i.e., hy-pokryal and rhithral), but also conveying different loads to the lower streams (i.e., metakryal and krenal).
The recession period is a further hot moment for benthic microbial life in the glacial stream as illustrated by pattern congruency of chlorophylla, bacterial biomass, exopolysac-charides and – to some extent – bacterial carbon produc-tion. Decreasing turbidity that enhances light availability, in-creased concentrations of NO3-N and other ions, and increas-ing channel stability all are factors favorable to in-stream primary production – an important contribution to microbial heterotrophs. These findings thus suggest predominantly en-dogenous processes to foster microbial life during recession, while exogenous processes may be more important during snowmelt.
Our results provide evidence towards the possible role of cations for microbial biofilms during snowmelt and particu-larly during recession. Notably calcium and magnesium are well known to enhance bacterial adhesion to substrates and to influence the tertiary structure of microbial exopolysac-charides (e.g. Geesey et al., 2001). We thus suspect that the copious production of exopolysaccharides during reces-sion and early winter flow is supported by the high avail-ability of calcium and magnesium. A well-developed ex-opolysaccharide matrix would translate into enhanced immo-bilization and storage of organic substrates and hence into an advantage during prolonged substrate deprivation (cf. Free-man and Lock, 1995; Battin et al., 1999). Furthermore, el-evated exopolysaccharide to cell ratios in the glacial stream (notably in the metakryal) where flow velocity is high hints at the function of exopolysaccharides providing better adhe-sion and protection from eroadhe-sion – a relationship that agrees with experimental findings (Battin et al., 2003). This advan-tageous style of microbial life in a copious matrix is also re-flected by elevated carbon production and by larger, more rod-like cells during late recession – a pattern that was par-ticularly noticeable in the glacial stream sites.
Spatial and temporal patterns of microbial cell carbon con-tent could also be attributable to resource dynamics. For in-stance, organic carbon of microbial/algal origin (as derived from fluorescence, McKnight et al., 2001) from glacial ice generally has better bioavailability than terrestrial organic carbon. This could explain the occurrence of larger cells in the metakryal during more favorable flow conditions. By contrast, smaller and coccoid-like cells during ablation hint at the glacial source of bacterial cells, which supports previ-ous findings (Battin et al., 2001). After a prolonged period of snow cover, however, average bacterial cell volume can be as low as 0.04µm3in the glacial stream (A. Wille, unpublished data) which may be a consequence of continuous starvation. These findings suggest that early winter could be important for microbial processes in the glacial stream, a notion that is in fact supported by observations from macroinvertebrate ecology (Sch¨utz et al., 2001).
Although water temperature is of general importance for bacterial growth in planktonic communities (Nedwell, 1999), our results did not detect any immediate effect of water tem-perature bacterial biomass and production. This may be at-tributable to the low temperature range (−0.1◦C to 9.8◦C)
for bacteria adapted to the life in the cold or to the syner-gistic effects of temperature and substrate supply. In fact, it is well known that microbial heterotrophs in cold aquatic ecosystems require high substrate concentrations to be active because of reduced substrate affinity due to stiffening of cell membrane lipids and decreased efficiency of the transport proteins (Nedwell, 1999). Thus, reduced substrate affinity at low temperature requires higher substrate concentrations to maintain cell metabolism. Relationships between bacte-rial production and both chlorophyllaand DOC fluorescence as descriptors for substrate availability, but also the relation-ship between bacterial production and EPS as a proxy for the biofilm storage capability (Freeman and Lock, 1995; Bat-tin et al., 1999) substantiate the overriding role of substrate availability versus temperature. This is in stark contrast to the effect of temperature for metazoan communities in alpine streams (F¨ureder et al., 2001; Milner et al., 2001).
Tributaries (rhrithral, krenal) were characterized by higher sediment organic matter, microbial biomass and bacterial carbon production. Substantially higher chlorophylla val-ues in the tributaries suggest authochtonous primary produc-tion as an important source of benthic organic matter, which is certainly subsidized from adjacent terrestrial vegetation. Lower scour in the tributaries contribute to the accumula-tion of this organic matter. Assuming a growth efficiency of 30% (Benner, 1988), the average theoretical turnover time of benthic organic carbon ranges from 196 to 226 d in the trib-utaries, but from 229 to 581 d in the glacial streams. These turnovers are reasonably close to values reported from tem-perate streams (Fischer et al., 2002, and references therein), which suggests efficient use of available organic carbon, and emphasizes the role of microbial biofilms in the carbon fluxes for high-alpine streams with otherwise low retention capac-ity. This complements recent work by Logues et al. (2004) showing the relationship between sediment organic matter and bacterial abundance in alpine streams.
In the context of alpine stream networks, rhithral and kre-nal streams may thus function as hotspots with dispropor-tionately high microbial and biogeochemical reaction rates (McClain et al., 2003) in an otherwise “cold” landscape. During increased hydrological linkage among streams – as indicated by the geochemical hysteresis curves – source-sink interactions between “hotspot” streams and channels with re-duced retention may contribute to ecosystem functioning at the network scale.
tem-plate and provides evidence that landscape filters (sensu Poff, 1997) can even act at the microbial scale. In fact, detec-tion rates of Eubacteria andα-Proteobacteria clearly seg-regated the main glacial stream from both tributaries, and this pattern even persisted after the unusually strong storm in September. Although this storm temporarily maximized among-stream linkage and increased the relative contribution ofArchaeaand members of theCytophaga-Flavobacterium cluster to biofilm communities in all study streams, the main glacial stream was still characterized by elevated numbers of theα-Proteobacteria. This supports recent molecular finger-printing that shows distinct bacterial community composi-tion in glacial and groundwater fed streams in the Swiss Alps (Logue et al., 2004). The higher detection rates of eubacterial cells in the tributaries likely reflect their elevated metabolic activity, a notion that is supported by the relationship be-tween cell metabolic rate and rRNA content (e.g., Bouvier and del Giorgio, 2003). The adaptation ofα-Proteobacteria to very low inorganic and organic nutrient-source concen-trations (Zavarzin et al., 1990) may render this group more competitive in the glacial stream.
In conclusion, our results have identified snowmelt and glacial ablation as important drivers in the structure and func-tion of microbial biofilms in alpine streams – both as hy-drologic drivers and sources of organic and inorganic sub-stances. Growing awareness of alpine streams as sensitive in-dicators for global change has recently fostered the develop-ment of various models predicting the structure of macroin-vertebrate communities in glacial streams (e.g., Smith et al., 2001; F¨ureder et al., 2001). Biofilms constitute a link be-tween hydrogeochemistry and benthic invertebrates, and we thus consider our findings as a stimulus to complement these models by explicitly integrating the coupling between terres-trial/aquatic connectivity, hydrogeochemistry and microbial ecology of alpine streams.
Acknowledgements. We thank W. M¨uller and J. Franzoi for
chem-ical analyses, M. Panzenb¨ock for field assistance and C. Hansen for providing the map in Fig. 1. The comments of two anonymous reviewers improved a former version of the manuscript. T. J. Battin was supported by the Jubil¨aumsfonds der ¨Osterreichischen Natio-nalbank (No. 7323), the Landesregierung Tirol and a Ramon y Cajal fellowship.
Edited by: J. Middelburg
References
Alfreider, A., Pernthaler, J., Amann, R., Sattler, B., Gl¨ockner, F. O., Wille, A., and Psenner, R.: Community analysis of the bacte-rial assemblages in the winter cover and pelagic layers of a high mountain lake by in situ hybridization, Appl. Environ. Micro-biol., 62, 2138–2144, 1996.
Battin, T. J., Butturini, A.,, and Sabater F.: Immobilization and metabolism of dissolved organic carbon by natural sediment
biofilms in a Mediterranean and temperate stream, Aquat. Mi-crob. Ecol., 19, 297–305, 1999.
Battin, T. J., Wille, A., Sattler, B., and Psenner, R.: Phylogenetic and functional heterogeneity of sediment biofilms along environ-mental gradients in a glacial stream, Appl. Environ. Microbiol., 67, 799-807, 2001.
Battin, T. J., Kaplan, L. A., Newbold, J. D., Cheng, X., and Hansen, C.: Effects of current velocity on the nascent architecture of stream microbial biofilms, Appl. Environ. Microbiol., 69, 5443– 5452, 2003.
Benner, R., Lay, J., K’nees, E., and Hodson, R. E.: Carbon conver-sion efficiency for bacterial growth on lignocellulose: implica-tions for detritus-based food webs, Limnol. Oceanogr., 33, 1514– 1526, 1988.
Bouvier, T. and del Giorgio, P. A.: Factors influencing the detec-tion of bacterial cells using fluorescence in situ hybridizadetec-tion (FISH): A quantitative review of published reports, FEMS Mi-crobiol. Ecol., 44, 3–15, 2003.
Boyer, E. W., Hornberger, G. M., Bencala, K. E., and McKnight, D. M.: Response characteristics of DOC flushing in an alpine catchment, Hydrol. Processes, 11, 1635–1647, 1997.
Brooks, P. D., McKnight, D. M., and Bencala, K. E.: The rela-tionship between soil heterotrophic activity, soil dissolved or-ganic carbon (DOC) leachate, and catchment-scale DOC export in headwater catchments, Wat. Resour. Res., 35, 1895–1902, 1999.
Brown, L. E., Hannah, D. M., and Milner, A. M.: Alpine stream habitat classification: an alternative approach incorporating the role of dynamic water source contributions, Arct. Antarct. Alp. Res., 35, 313–322, 2003.
Burgherr, P., Ward, J. V., and Robinson, C. T.: Seasonal variation in zoobenthos across habitat gradients in an alpine glacial flood-plain (Val Roseg, Swiss Alps), J. North Am. Benthol. Soc., 21, 561–575, 2002.
Campbell, D. H., Clow, D. W., Ingersoll, G. P., Mast, M. A., Spahr, N. E., and Turk, J. T.: Processes controlling the chemistry of two snowmelt-dominated streams in the Rocky Mountains, Wat. Resour. Res., 31, 2811–2821, 1995.
Chapin III, F. S. and K¨orner, C.: Arctic and alpine biodiversity: pat-terns, causes and ecosystem consequences, Trends Ecol. Evol., 9, 45–47, 1994.
Dubois, M. G., Hamilton, K. A., Rebers, J. K., and Smith, R.: Col-orimetric method for the determination of sugars and related sub-stances, Anal. Chem., 28, 350–356, 1956.
Engstrom, D. R., Fritz, S. G., Almendinger, A. E., and Juggins, S.: Chemical and biological trends during lake evolution in recently deglaciated terrain, Nature, 408, 161–166, 2000.
Fischer, H., Wanner, S. C., and Pusch, M.: Bacterial abundance and production in river sediments as related to the biochemical com-position of particulate organic matter (POM), Biogeochemistry, 61, 37–55, 2002.
Freeman, C. and Lock, M. A.: The biofilm polysaccharide matrix: A buffer against changing organic substrate supply?, Limnol. Oceanogr., 40, 273–278, 1995.
F¨ureder, L., Sch¨utz, C., Wallinger, M., and Burger, R.: Physico-chemistry and aquatic insects of a glacier-fed and a spring-fed alpine stream, Freshwat. Biol., 46, 1673–1690, 2001.
bac-teria and diatoms: A review, Biofouling, 15, 195–205, 2000. Hieber, M., Robinson, C. T., Rushford, S. R., and Uehlinger, U.:
Algal communities associated with different alpine stream types, Arct. Antarct. Alp. Res., 33, 447–456, 2001.
Hood, E. W., McKnight, D. M., and Williams, M. W.: Sources and chemical character of dissolved carbon across an alpine/subalpine ecotone, Green Lakes Valley, Colorado Front Range, United States, Wat. Resour. Res., 39, 1188–2000, 2003. Kaufmann, R.: Invertebrate succession on an alpine glacier
fore-land, Ecology, 82, 2261–2278, 2001.
Krebs, C. J.: Ecological methodology, Harper Collins Publishers, New York, 654, 1989.
Liu, F., Wiliams, M. W., and Caine, N.: Source water and flow paths in an alpine catchment, Colorado Front Range, United States, Wat. Resour. Res., 40, W09401, doi:10.1029/2004WR003076, 2004.
Logue, J. B., Robinson, C. T., Meier, C., and Van der Meer, J. R.: Relationship between sediment organic matter, bacteria compo-sition, and the ecosystem metabolism of alpine streams, Limnol. Oceanogr., 49, 2001–2010, 2004.
Malard, F., Tockner, K., and Ward, J. V.: Shifting dominance of sub-catchment water sources and flow paths in a glacial floodplain, Val Roseg, Switzerland, Arct. Antarct. Alp. Res., 31, 135–150, 1999.
Malard, F., Tockner, K., and Ward, J. V.: A landscape-level anal-ysis of physico-chemical heterogeneity in a glacial floodplain, Landsc. Ecol., 15, 679–695, 2000.
Marxsen, J.: Measurement of bacterial production in stream-bed sediments via leucine incorporation, FEMS Microbiol. Ecol., 21, 313–325, 1996.
Max, L. and Onghena, P.: Some issues in the statistical analysis of completely randomized and repeated measures designs for speech, language and hearing research, J. Speech Lang. Hear. Res., 42, 261–270, 1999.
McClain, M. E., Boyer, E. W., Dent, C. L., Gergel, S. E., Grimm, N. B., Groffman, P. M., Hart, S. C., Harvey, J. W., Johnston, C. A., Mayorga, E., McDowell, W. H., and Pinhay, G.: Biogeochemi-cal hotspots and hot moments at the interface of terrestrial and aquatic ecosystems, Ecosystems, 6, 301–312, 2003.
McGregor, G., Petts, G. E., Gurnell, A. M., and Milner, A. M.: Sen-sitivity of alpine stream ecosystems to climate change and human impacts, Aquat. Conserv. Mar. Freshwat. Ecosyst., 5, 233–247, 1995.
McKnight, D. M., Boyer, E. W., Westerhoff, P. K., Doran, P. T., Kulbe, T., and Andersen, D. T.: Spectrofluorometric character-ization of dissolved organic matter for indication of precursor organic material and aromaticity, Limnol. Oceanogr., 46, 38–48, 2001.
Milner, A. M., Brittain, J. E., Castella, E., and Petts, G. E.: Trends in macroinvertebrates community structure in glacier-fed rivers in realtion to environmental conditions: a synthesis, Freshwat. Biol., 46, 1833–1848, 2001.
Nedwell, D. B.: Effect of temperature on microbial growth: lowered affinity for substrates limits grwoth at low temperature, FEMS Microbiol. Ecol., 30, 101–111, 1999.
Norland, S.: The relationship between biomass and volume of bac-teria, In: Handbook of methods of aquatic microbial ecology, edited by: Kemp, P. F., Sherr, B., Sherr, E., and Cole, J. J., Lewis Publ. BocaRaton, 303–308, 1993.
Peterson, C. G., Valett, H. M., and Dahm, C. N.: Shifts in habitat template for lotic microalgae linked to interannual variation in snowmelt intensity, Limnol. Oceanogr., 46, 858–870, 2001. Poff, N. L.: Landscape filters and species traits: towards a
mecha-nistic understanding and prediction in stream ecology, J. N. Am. Benthol. Soc., 16, 391–409, 1997.
Porter, K. G. and Feig, Y. G.: The use of DAPI for identifying and counting aquatic microflora, Limnol. Oceanogr., 25, 943–948, 1980.
Posch, T., Pernthaler, J., Alfreider, A., and Psenner, R.: Cell-specific respiratory activity of aquatic bacteria studied with the tetrazolium reduction method, cyto-clear slides, and image anal-ysis, Appl. Environ. Microbiol., 63, 867–873, 1997.
Psenner, R. and Nickus, U.: Snow chemistry of a glacier in the Central Eastern Alps (Hintereisferner, Tyrol, Austria), Zeitschr. Gletschk. Glaziol., 22, 1–18, 1986.
Sattler, B., Puxbaum, H., and Psenner, R.: Bacterial growth in supercooled cloud droplets, Geophys. Res. Lett., 28, 239–242, 2001.
Sch¨utz, C., Wallinger, M., Burger, R., and F¨ureder, L.: Effects of snow cover on the benthic fauna in a glacier-fed stream, Fresh-wat. Biol., 46, 1691–1704, 2001.
Smith, B. P. G., Hannah, D. M., Gurnell, A. M., and Petts, G. E.: A hydrogeomorphological context for ecological research on alpine glacial rivers, Freshwat. Biol., 46, 1579–1596, 2001.
S¨awstr¨om, C., Mumford, P. Marshall, W., Hodson, A., and Laybourn-Parry, J.: The microbial communities and primary productivity of cryconite holes on an Arctic glacier (Svalbard 75◦N), Pol. Biol., 25, 591–596, 2002.
Tockner, K., Malard, F., Burgherr, P., Robinson, C. T., Uehlinger, U., Zah, R., and Ward, J. V.: Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland), Arch. Hydrobiol., 140, 433–463, 1997.
Tockner, K., Malard, F., Uehlinger, U., and Ward, J. V.: Nutri-ents and organic matter in a glacial river-floodplain system (Val Roseg, Switzerland), Limnol. Oceanogr., 47, 266–277, 2002. Tscherko, D., Rustemeier, J., Richter, A., Wanek, W., and Kandeler,
E.: Functional diversity of the soil microflora in primary succes-sion along two glacier forelands in the Central Alps, Eur. J. Soil Sci., 54, 685–696, 2003.
Velji, M. I. and Albright, L. J.: Microscopic enumeration of at-tached marine bacteria of seawater, marine sediment, fecal mat-ter, and kelp blade samples following pyrophosphate and ultra-sound treatments, Can. J. Microbiol., 32, 121–126, 1985. Ward, J. V.: Ecology of alpine streams, Freshwat. Biol., 32, 277–
294, 1994.
Williams, M. W., Hood, E. W., and Caine, N.: The role of organic nitrogen in the nitrogen cycle of a high-elevation catchment, Col-orado Front Range, USA, Wat. Resour. Res., 37, 2569–2582, 2001.
